The present paper forms the first part of a contribution to a comprehensive critical evaluation of the use of pharmaco-electroencephalography (pharmaco-EEG) as an instrument in the evaluation of pharmacodynamic effects of psychoactive drugs in man and further prediction of therapeutic effects by utilizing partial effects in a clinical pharmacological model situation. The basic principles and the methodological approaches are described first, prior to discussing the possibilities and limitations of the pharmaco-EEG in postulating hypotheses about therapeutic efficacy of psychoactive drugs in man. The discussion rests upon the findings with three selected new psychoactive drugs (savoxepine, levoprotiline and maroxepine), presented in this (Part I) and subsequent (Part II and III) reports. In this report results obtained with savoxepine, a novel tetracyclic compound, are described. Convergent evidence for antipsychotic potential of this drug has emerged from preclinical and pharmaco-EEG studies, subsequently corroborated by the first open trials in patients suffering from acute episodes of schizophrenia. In the pharmaco-EEG study a single dose of 0.5 mg savoxepine was tested in 15 healthy subjects with comparison to chlorpromazine (75 mg) and placebo. The experimental design was a threefold cross-over uncompleted block design with three consecutive trial days one week apart. Pharmaco-EEG was recorded under a high situational level of vigilance before 3 h and 6 h after drug administration. The recording was performed by means of 12 electrodes placed in the frontal, central, parietal and occipital regions. EEG-maps were constructed after power spectral analysis of absolute power changes in predetermined frequency bands within the range 0.5 to 30.0 Hz. The findings indicated EEG effects of savoxepine similar to those of chlorpromazine. In the tested dosage, however, savoxepine showed higher potency, later maximum (6 h) of the effect and lower sedative potential. Clinical trials performed in two centres and in, altogether, 28 acutely psychotic patients indicated antipsychotic activity of the drug in a low range of doses (on average, 0.1 to 0.9 mg/d). Even though the antipsychotic efficacy of savoxepine remains to be confirmed in controlled studies, the results favour the hypothesis that pharmaco-EEG has a value in predicting the quality of action of a new drug in man. Results also favour the complementary use of pharmaco-EEG in human pharmacology testing in order to objectively assess central effects of drugs and rationally estimate the doses for therapeutic trials.